Forget Intel Broadwell, Skylake On the Way

16GB DDR3 SO-UDIMM does exist, Intel blocked it though.

 

Really? Where did u read that Intel blocked it?

@Charles P. Jefferies, I'm not saying that Skylake won't be an improvement but the thing is how much are we going to wait for it? Broadwell has slipped so badly and honestly, I don't expect any high performance Skylake parts in mobile/desktop till 2016 tbh even if Intel say some desktop parts will be available in H2 2015... I can afford to wait since I've literally maxed out my 3940XM and 970M but not everyone can... Also, the emphasis on GPU with broadwell recently and not much improvements CPU side is a bit irritating for me since really iGPU's are good enough for me once my requirements above have been met which they almost have... Well that's just me..

 

Can't remember the source. Will search for it when I have time.

When the Haswell platform (or anything earlier) was released, 16GB DDR3 UDIMM didn't exist. Since it's impossible to test with real hardware Intel deliberately disabled the support even though the capacity is allowed by the DDR3 spec. After 16GB sticks came out some people managed to hack it and it worked. It's a trivial software tweak and some mobo vender even rolled out BIOS updates to support it officially.

AMD didn't disable anything that follows the spec and 16GB UDIMMs work well out of the box on AMD platforms. Not that I want to use a AMD laptop though.

 

Skylake uses 60 percent lower power

Skylake Notebook coming in Q4 2015

 

^Drooools
What can't Skylake do?

 

Live up to its hype

 

Q4 2015.. Hmm.. Let's wait and see.. I'm more interested in the Clevo laptops with desktop processors so hopefully skylake makes it in those laptops...

 

At this stage in the cycle of everything, I am looking forward to my next notebook having DDR4 as standard, which will mean no doubt Skylake. This thread was started a long time ago, but its title is actual starting to be relevant about now.

 

http://www.digitaltrends.com/computing/intel-skylake-processors-tipped-to-debut-in-august-not-june/

 

Intel Preparing Something ‘Big’ and ‘Disruptive’ with Skylake Microarchitecture

 

This article was interesting to me with regards to the delays people are always lamenting about here.

See:
http://www.anandtech.com/show/8991/...ng-the-benefits-of-14nm-and-going-beyond-10nm

Some of the comments were equally enlightening too.

 

Mainstream Intel Core Processors will not support AVX 512 from Skylake – Only Xeon

 

Skylake looks more interesting every day.. It looks like it's going to be well worth the wait!

 

That is what they always say....

So Skylake is going to be huge because Intel won't say anything about it? That fails logic.

 

Nobody really said that about Ivy, Haswell, or Broadwell, outside of the GPU.

 

Really? I knew quite a few people who thought Haswell was going to be another Sandy Bridge in terms of CPU performance improvement and overclockability, only to be floored when reality hit.

 

Intel's Skylake chips to appear in tablets, PCs, servers
Intel CEO Brian Krzanich says partners, including PC makers, are looking forward to Skylake

 

What are these laptops that Dell and HP are supposedly skipping Broadwell with? I can think of Dell Latitude E6xxx for one, but are they really going to wait that long to update it when the standard voltage BDW chips are right around the corner?

 

I think we just assumed, because it was a tock, right? Then when we got closer to release, we began to hear that IPC wasn't going to be much. Close to release, Skylake hasn't fit that model. Plus we know there are some concrete platform upgrades that will be pretty worthwhile.

 

I think that statement would only apply to the Broadwell quad-cores which have yet to appear for notebooks - but that would encompass more than just Dell since everyone is at the mercy of Intel. As a matter of fact, does anyone remember where Dell officially (or unofficially) said they were skipping Broadwell for some notebooks? I would like to see the source.

 

PC World had this article: Dell skips Broadwell with some laptops, plans for Skylake where they said:

That is the closest official statement I have seen.

 

That is old, but yea about the only thing around.

Get the feeling that might be a very long wait for Skylake. Might be hard to ignore Broadwell if all there we have got it just recycled Haswell parts.

 

See:
http://forum.notebookreview.com/thr...kylake-on-the-way.724059/page-14#post-9937200

 

see what?

 

Why I'm waiting for Skylake:

• USB type-C is more likely to be included
• NVMe SSDs are more likely to be included
• IGZO IPS panels are more likely to be included
  
• LPDDR3 is required on the Y- and U-series (substantial battery savings over DDR3L)
  
• SKL-U (Ultrabooks) will have the PCH integrated onto the die (packages possibly smaller than BDW-Y/Core M)
• Another generation for OEMs to iterate (look at the new XPS 13, the Lenovo HZ750, etc.)

For that last point, I really wish we had a premium PC OEM. None of these half-baked touchpads, 720p webcams, crappy wireless performance, hair-pulling warranties and support, gimped battery life from an excessively high DPI screen, overheating concerns, etc. I can't believe we're on the 5th generation of Intel CPUs and we still deal with all this garbage.

Bring me an amazing $2000 laptop. I will pay.

--

Some other nice things about Skylake, but not worth waiting for:

• Rezence wireless charging support from "every major PC vendor", though probably just select models
• Windows 10 installed out-of-the-box
• Intel HD 6000 series will support Direct3D 12.0 at feature-level 12.0

 

Halleluiah! Bring on amazing. (And kill the 'magical' witch while we're at it).

(If the rumors are true, don't have to wait for Win10 - if the free upgrade is true for Win7 and up systems).

 

I believe it has been confirmed to be a free upgrade tiller for all 7/8.1 owners for at least 1 year from launch I believe..

 

how much more of a performance boost are we looking at? i heard intel say it will be the biggest performance boost in over a decade. so 10 percent? 1000 percent?

i'm interested to see if it's worth it to go from haswell to skylake for video rendering

 

Don't they say this every time...

If it really is a bigger change than SNB that would be good. But can we trust them?

 

So your reasons for waiting for Skylake are for things that have literally nothing to do with Skylake? And who cares about LPDDR3? Squeezing a tiny bit more power savings out of something that already uses almost no power doesn't help and integrating the PCH doesn't change anything unless you design motherboards for a living.

 

AFAIK LPDDR3 is already compatible with some/all existing Haswell platforms, its just really expensive, and not worth the extra $$$ for the minimum gain in battery life.

 

When is mobile Skylake ala i7 quad core (or possibly 6 to 8 core) supposed to come out?
If its not until next year, then I think there's no point in waiting for that long

 

I can't speculate why we haven't had a true premium PC OEM for laptops. Sony was, in my opinion, the closest, but even they cut corners. It's like I'm willing pay for a Rolls-Royce, but the only options I have are millions of Pintos, a few thousand Camry's, and a handful of Beetles (plus one or two Formula 1 Racecars..the crazy X99 laptops).

Not a rumor any longer: Win10 will be free for all Windows 7 and up users (for the first year after Win10 releases).

But, when I wrote that post, I suspected there may be some hardware needed to get all of Windows 10's features. And turns out that's at least true for the new infrared-based facial recognition system (called "Windows Hello"). From the linked Ars Technica article,

If Intel has only made incremental progress from 2011 to 2015 (Sandy Bridge to Broadwell), I'm assuming we are in for another round of "incremental" CPU performance increases.

If you have Haswell now...just hold on for the reviews in a few months to make a comparison.

Exactly. Expect little. If we get a nice surprise, good for us.

Yep. CPU performance in general consumer uses hit a plateau years ago. I pity the fool who is waiting expecting anything besides 5-10% CPU performance increases with Skylake. It's like being Apple and feeling ecstatic that USPS says it will deliver packages 5% faster next year, lol. Hooray?

LPDDR3: see below.

"unless you design motherboards for a living" = well, I sure am happy that every laptop PC OEM designs their own motherboards! lol, what....

Right, Haswell supports LPDDR3 and DDR3L; OEMs just choose.

LPDDR3 uses 19% less power than DDR3L in active use, but in standby mode, it uses almost 9x less power than DDR3L. 9x less. But, it's just RAM, right? How much power is it using, anyways? Well, Samsung claims its LPDDR3 shows 5 more hours of standby than DDR3L. More information from Intel here.

It's not "really expensive". 4GB DDR3L costs $28, while 4GB LPDDR3 costs $36. $8 (or $16 if 8GB) for 5 more hours of standby power? No brainer.

There's a reason that the high-end laptops use LPDDR3:

Macbook Air (Haswell and Broadwell)
2015 Macbook
2015 Macboo Pro
Surface Pro 3
Sony Vaio Pro 13
Acer Aspire R 13
etc.

 

Any idea about when Skylake H (47W) is expected to come out? And will Alienware go straight to Skylake or wait go to Broadwell in the mean time?

 

Standby is measured in days, not hours. 19% less power in active use isn't a whole lot when we are talking about the RAM using 1 watt.

 

We have X99 laptops now?

 

Where Where? I only know of X79 and Z97 Clevo laptops

 

Your use of the terms 'standby' & 'active' are somewhat vague.

Let's look at some real figures/values:

According to Micron, for a 4GB stick of DDR3L RAM, the typical power used during ' active-standby' (i.e. the RAM is idling, no read/writes are occuring, but RAM is refreshing) is approximately 100mW or 0.1W. For 2 modules (a more common configuration), this might result in 0.2W. For your claimed reduction of 20%, this would result in a 0.04W saving.

For a heavier usage scenario of 'writes', power consumption is typically 0.4W. Again, for 2 modules, this might be 0.8W. A 20% claimed saving would be around 0.16W.

Of course every saving counts, but at a time where processors are still using approx 1.5W during idle and current displays around 4W when idle, many would call these insignificant savings. And this is only a small part of the picture. (Ignoring other power losses/inefficiencies, DC-DC conversions, other component power draws, storage, network card, fans, etc.)

EDIT: Source here: https://www.google.co.uk/url?sa=t&r...=_XB3i4lDrQawzt4Z0seoUw&bvm=bv.88528373,d.ZGU

 

Am I the only guy who thinks the most significant power saving in the near future could come from LED displays and white-text-on-black-background colour schemes?

 

Don't want to go back to orange/two 'colour' CRT displays, but I do agree that the display is the next challenge for manufacturers to lower the overall power draw.

After Skylake arrives and all it promises, whether that display revolution will be LED based or not remains to be seen, but whatever brainpower is put towards it will be more than worth it if the platform power otherwise allows for all day+ computing... but at one quarter to one half the display brightness (currently).

 

Well, let's just settle this LPDDR3 hoopla now. I found some more data; it's based more on tablets and other mobile hardware, but I Microsoft is pushing laptop/tablets in that same "Always On, Always Connected" principle.


From Micron. Granted, it's from a tablet and not a laptop, but those categories are getting closer and closer. 44 days increase in standby time by switching to LPDDR3 from DDR3L.

I'm not concerned with active RAM use; just when it's in standby (exactly: just DRAM refreshes to maintain the data, nothing else). LPDDR3 has PASR (partial array self-refresh); a feature missing on DDR3/DDR3L. A primer on DRAM Refresh & PASR from Microsoft.

From DRAM paper, page 116.

From DRAM paper 2, first page.

I'm no RAM or power efficiency expert, but all of this data + Intel's requirement for LPDDR3 on Skylake-U + the presence of LPDDR3 on most high-end laptops leads me to believe LPDDR3 has some tangible benefits for laptop standby battery life. And, it'll have nearly price parity with DDR3L in the coming months, anyways. If you still don't care about LPDDR3, fine, go back to your house under a rock, heh. I'm excited for it.

Whoooops, sorry. I read an article about LGA2011 laptops; I forgot X99 is LGA2011- v3. My bad.

IGZO-based LCDs are likely the next part of the display power use revolution. Its thinner transistors can reduce backlight usage (the big power draw) because you don't have all these big transistors blocking out light.

From DisplayMate. I guess IGZO was still maturing at the time; now we've finally got 13" models coming out...

IGZO backplates with OLED: the dream...

 

Why, hello there, Skylake... From the minimum hardware requirements for Windows 10 posted on March 18th. They call Skylake an SoC...I wonder if they're referring to the Y-series, or now also the U-series, seeing as we really are just one big die now, heh.

 

As Qing Dao mentioned, and as your figures state, standby time is measured in days. Yes the increase from 11 days to 55 days is a very long one, but may I ask you a question:

If you asked the average laptop user what was the improvement they wanted most on their device - do you think their answer would be a 55 day standby time? Nup, I didn't think so either.

The overwhelming majority of people would be eager to have noticeably longer battery life when actually USING the device, and doing something useful with it. As mentioned by several posters, things like long-awaited improvements in LED display technology will help achieve this, as well as new battery technology; not a measly 20% reduction in active mode RAM power consumption.

There is no doubt that LPDDR3 is yet another step in moving forward, but you seem to think it magically solves all the issues & demands of lower power consumption.

You also should learn the phrase before you use it: It is actually: "go back to living under a rock". Keep that in mind.

 

All I said was I am excited for LPDDR3 coming to Skylake and that "believe LPDDR3 has some tangible benefits for laptop standby battery life". I never said it solves all issues and demands of lower power consumption: where did you read that?

I think you're reading some different posts than I am.

Look at the cost-benefit analysis. If you asked the average laptop user would they pay $18-36 for 5x the standby battery life, would they do it? Yes, I think most would.

But, do you know the best part? With Skylake Ultrabooks, that's not even a choice anymore: everyone gets 55 day standby time! Er, more or less, haha.

Bahahaha, I know the phrase very well, thank you, however. I just wanted to make it sound a little more interesting because "go back to living under a rock" is a little trite these days. And I want to make the allusion that not only do you live under a rock, but you enjoy it so much that you've built a house. There's a little garden in the front-yard, a school down the street. This is your actual home.

---

I am surprised my post caused such controversy. All I mentioned was that LPDDR3 is coming to Skylake and I'm excited for it. And the response?

"How DARE you be excited for that? Get out of here."
"My antiquated/incorrect knowledge is correct because it is commonly accepted! Come, friends, jump on this bandwagon but don't search on Google"
"His claim is unusual; it must be wrong!"

It's like everyone got excited to pounce on a guy for making some outlandish claim....and then got #rekt when he dropped some data.

Anyways, I'm sure everyone else has learned something about DRAM, standby power consumption, LPDDR3, and upcoming pricing changes.....while you and Qing Dao go on about something or another,

 

Cost-benefit analysis you say? Ok.

Cost: "$18-36"

Who can benefit from this: A person who uses their laptop once or twice a month, does not want to plug it in, and can't stand to wait a few extra seconds for their laptop to start from hibernation.

Analysis: An $18-36 surcharge for everybody, benefit to at best 1% of laptop users. Waste of money if given the choice.

I really doubt that many people would choose the option for LPDDR3 after weighing the facts.

If you want to be excited about not using your computer for almost two months, be my guest.

You act like LPDDR3 actually means anything. It doesn't. Just like how DDR2-DDR3 didn't change anything or DDR3-DDR3L didn't change anything either. You and the other ten people in the world who use their laptops once a month and don't want to hibernate or plug it in can rejoice. For everyone else, DDR3L-LPDDR3 is just another big yawn.

This can definitely help with tablets, but Intel forcing "ultrabooks" to use LPDDR3 helps nearly nobody besides makers of LPDDR3. If there was such a big deal with this stuff, more laptop manufacturers would have actually wanted to use LPDDR3 in their Haswell and Broadwell laptops instead of waiting for Intel to force them to.

 

Nobody really know it.
Rumors talks about going directly to H Skylake on 3Q release (August IDF) bypassing Broadwell, with the lacks of broadwell parts (Xeon, Desktop, H Mobile) make sense.
The 65W K part showed recently is pretty strange.... and it's a niche products BTW.

Something is really wrong with Broadwell/14nm. Launched at last Sept IDF so far only Core M and U parts.... even Braswell is delayed to Mid 2015, or maybe Broadwell is simply killed.
The Core M and U parts are the CPUs with more competitions by ARM with Tablets, everything else is pretty much unchallenged by far.
The question is why rush Skylake on the market when the 10nm is running later, so that Cannonlake is a 2017 parts at the earliest.
With Suppose Skylake-K on the market in september, the sales of X99 and Haswell-E will go even lower.. 2 generations ahead, cheaper and faster, and support at 64GB, only the $1000 X will make sense..

 

Originally, you didn't specifically mention 'standby battery life', you only said 'battery life'.

You also posted a claim stating that LPDDR3 RAM has 20% lower active power consumption, compared to DDR3L.

All I ever did was produce & analyse some detailed facts and figures, which put your claims into perspective, and showed that a 20% power consumption reduction (for active usage) is far less significant than it actually seems.

How dare I NOT get excited for LPDDR3!

LOL, how old are you? Name one instance of a claim I made, which you proved wrong with facts and figures? Nope, cause it never happened.

Anyway, have a nice weekend, and I hope you enjoy your LPDDR3 RAM too by the way, and the HUGE difference it will make in day to day usage.

 

Who cares about DDR3 of any kind, just give me DDR4 in a notebook, just like I use in my desktop currently. And all the benifits that will come with Skylake.

This new processor and chipset cannot get here fast enough, just about purchased another Hasbeen (I mean Haswell) notebook. Lucky I cancelled before it was to late, other would be stuck with a Windows 8.1 with processor and chipset from way back in 2013!

Obviously Broadwell mobile quad has been forgotten now we are just a few days from the start of Q2. Bring on Skylake. I am guessing Computex first week of June we will see something.

 

Broadwell quad core CPU might never come to the market. We are at the end of 1Q and all we have are i5 5200u and i7 5500u. and they are at a tie with their haswell i5 and i7 U variants except better integrated GPU performance. They can drive 4k using single DP cord.

 

So, if I'm not really in a rush but anyway really eager to buy a laptop. Is it worth the wait for broadwell/skylake? Or am I better off just buy one now with Haswell? It's supposed to be a 15" that should be capable of gaming and photo editing. Price about 1500$.

I do need one at the latest in August as that's when the most photo work is going to start again.